-
ED 475 633
AUTHOR
TITLE
INSTITUTION
SPONS AGENCYREPORT NOPUB DATENOTE
CONTRACTAVAILABLE FROM
PUB TYPE
EDRS PRICE-
DESCRIPTORS
IDENTIFIERS
ABSTRACT
DOCUMENT RESUME
PS 031 232
Gordon, Anne; Briefel, Ronette; Needels, Karen; Wemmerus,Nancy;
Zavitsky, Teresa; Rosso, Randy; Tasse, Tania; Kalb,Laura; Peterson,
Anne; Creel, Darryl
Feeding Low-Income Children When School Is Out: The SummerFood
Service Program. Final Report.Mathematica Policy Research,
Princeton, NJ.Economic Research Service (USDA), Washington,
DC.No-8790-0912003-03-00243p.
53-3K01-1-7For full text:
http://www.mathematica-mpr.com/PDFs/feedinglow.pdf.
Reports Evaluative (142)
EDRS Price MF01/PC10 Plus Postage.*Children; Low Income;
*Nutrition; Program Descriptions;Program Effectiveness; Program
Improvement; State Programs;Summer Programs
Sponsors; *Summer Food Service Program
The primary goal of the Summer Food Service Program (SFSP) isto
provide nutritious meals to children in low-income areas when
school isnot in session. This report presents the findings of the
SFSP ImplementationStudy, a descriptive study of the operations of
the SFSP at the state andlocal levels and how SFSP staff feel the
program could be improved. Thereport's introductory chapter
describes the SFSP program and its history, thestudy objectives and
research questions, the study design and data sources,and the plan
of the report. Chapter 2 presents a statistical portrait of theSFSP
in 2001 that covers the characteristics of the state agencies,
sponsors,and sites that operated the program. The chapter also
describes the SFSPparticipants and discusses factors that may
affect their participation. Eachof the next three chapters
addresses one of three research questions. Chapter3 describes how
the program was administered at the state, sponsor, and sitelevels,
as well as areas of program operations that staff believed could
beimproved. Chapter 4 discusses participation and outreach,
including staffviews on barriers to participation, the extent of
outreach efforts, and thereasons why sponsors left the program.
Chapter 5 describes the meals servedat the SFSP, their nutrient
content, and the extent of plate waste. FinallyChapter 6 summarizes
the findings and considers issues for the future. Amongthe findings
are the following: (1) around 70 percent of sponsors expectedthat
SFSP reimbursements would not cover all their costs; (2) all
stateagencies reported conducting outreach to attract new
participants; (3) staffat all levels most frequently cited lack of
transportation as a barrier tochildren's participation; (4) on
average, SFSP meals provided a least one-quarter of the Recommended
Daily Allowances for most key nutrients atbreakfast, and at least
one-third for energy and key nutrients at lunch andsupper; and (5)
children wasted an average of about one-third of the caloriesand
nutrients they were served. (Contains 40 references.) (HTH)
Reproductions supplied by EDRS are the best that can be madefrom
the original document.
-
Contract No.: 53-3K01-1-7MPR Reference No.: 8790-091
U.S. DEPARTMENT OF EDUCATIONOffice of Educational Research and
Improvement
EDUCATIONAL RESOURCES INFORMATIONCENTER (ERIC)
This document has been reproduced asreceived from the person or
organizationoriginating it.
1:1' Minor changes have been made toimprove reproduction
quality.
Points of view or opinions stated in thisdocument do not
necessarily representofficial OERI position or policy.
Submitted to:a.
Feeding Low-IncomeChildren When School IsOut: The Summer
FoodService Program
Final Report
March 2003
Anne GordonRonette BriefelKaren NeedelsNancy WemmerusTeresa
ZavitskyRandy RossoTania TasseLaura KalbAnne PetersonDarryl
Creel
USDA, ERS1800 M Street, NWRoom 3063 North TowerWashington, DC
20036-5831
Project Officer:Jane E. Allshouse
2
MATHEMATICAPolicy Research, Inc.
PERMISSION TO REPRODUCE ANDDISSEMINATE THIS MATERIAL HAS
BEEN GRANTED BY
1
TO THE EDUCATIONAL RESOURCESINFORMATION CENTER (ERIC)
Submitted by:
Mathematica Policy Research, Inc.P.O. Box 2393Princeton, NJ
08543-2393(609) 799-3535
Project Director:Anne Gordon
BEST COPY AVAILABLE
-
ACKNOWLEDGEMENTS
Many at USDA made important contributions to this study. We
would particularly like toacknowledge the helpful feedback and
support that Jane Allshouse, our project officer at theEconomic
Research Service (ERS), provided throughout all stages of the
project. Others at ERSwho helped guide the study design and who
commented on drafts of the report include DavidSmallwood, Betsy
Frazao, and Joanne Guthrie. Staff at the Food and Nutrition Service
(FNS)also made important contributions to the project. Anita Singh,
from the FNS Office of Analysis,Nutrition, and Evaluation, and
Linda Jupin, from the Child Nutrition Division, gave us
feedbackthroughout the study. Linda Jupin also answered many
technical questions about program rulesor helped direct us to other
sources. Susan Fouts and John Endahl, both of FNS, made
valuablecomments on the draft reports. Two other FNS staff, Joan
Tressler and Jeff Derr, provided keyprogram data. We also are
grateful to the following FNS regional office staff, who
helpedpromote the study to state agencies and sponsors, and helped
us resolve problems with the datacollection process: Sharon Bevan,
Alan Crane, Dick Gilbert, Mary Gioiosa, John Hanna,Doraetta Mozon,
and Ellen Wahlberg.
We want to express our appreciation to the staff at SFSP state
agencies, local sponsors,former sponsors, and sites who responded
to the survey questions, provided lists of sponsors andsites, and
permitted us to observe SFSP operations at local sites. Despite the
substantialdemands of participating in this study, all these
individuals were enormously helpful andcooperative.
Many hardworking, dedicated staff supported MPR's data
collection efforts, which were ledby Laura Kalb. Milena Rosenblum,
Laura's assistant throughout the project, was a key factor inthe
study's success. Milena worked on every phase of the data
collection, from initial calls tostates for sample lists, to
quality assurance review for the sponsor surveys, to conducting
manystate interviews, to coordinating collection of lists for the
Sponsor-Site Database. DonnaMikolajewski monitored project costs
throughout the project.
The sponsor survey was directed by Lindsay Crozier, with
assistance from Barbara Schiff.Barbara also helped collect sample
frame information, planned the field training, conducted
stateinterviews, and monitored the survey of former sponsors.
Theresa Boujada supervised telephonecenter operations for the
surveys of sponsors and former sponsors. Bea Jones developed
theAccess databases that were key to tracking data collection for
the study and the Sponsor-SiteDatabase that was delivered to ERS.
Bea also helped collect sample frame information andconducted state
interviews. Neil DeLeon, Mark Dentini, Linda Gentzik, and Ron
Palancaprovided programming support for the surveys. Sue Golden and
her staff performed yeomanwork in data entry of sample frame lists,
survey instruments, and, finally, lists of all 2001sponsors and
sites (close to 40,000 records).
Francene Barbour coordinated field efforts for the site visits,
with help from Mike Haas,Adrienne Wilschek, and LaShona Burkes.
They were responsible for sending a staff of 15 fieldinterviewers
to 38 states, helping them with travel arrangements, and dealing
with any problemsthat arose. (Adrienne and LaShona also helped with
the Sponsor Survey, and Adrienne later
iii
. 3
-
conducted state interviews.) At Garcia Research Associates, a
subcontractor to MPR, SteveGoodwillie coordinated Garcia's field
interviewers, who handled California site visits.
The sampling efforts for the study were led by Anne Peterson,
who built on design work byMichael Sinclair. Brenda Cox and Daniel
Kasprzyk provided quality assurance review ofsampling work. Miki
Satake provided programming support in selecting the samples, and
DarrylCreel programmed and documented the weights.
Nutritional coding of the data was conducted by Kate Gilstad and
Jennifer Cohen, under thesupervision of Ronette Briefel and Teresa
Zavitsky.
Analysis of the data involved a large team, in addition to the
authors. Jim Ohls reviewed thereport and provided valuable advice
throughout the study. Vatsala Karwe, Daisy Ewell, LauraFolks,
Joshua Hart, Jocelyn Lewis, Lucy Lu, and Sherry McDonald provided
programmingsupport. Laura Berenson edited the report, greatly
contributing to its readability. WilliamGarrett coordinated report
production.
iv
-
I. INTRODUCTION
The primary goal of the Summer Food Service Program (SFSP) is to
provide nutritiousmeals to children in low-income areas when school
is not in session. The program is intended tofill a gap in services
for low-income, school-age children, who receive free or
reduced-priceschool meals during the school year through the
National School Lunch Program (NSLP) and theSchool Breakfast
Program (SBP). However, the SFSP always has served far fewer
children thanare reached by free or reduced-price NSLP lunches
during the school year.
Because the SFSP, which had expenditures of $272 million in
fiscal year (FY) 2001, is oneof the smaller child nutrition
programs administered by the U.S. Department of Agriculture(USDA),
it has not been studied extensively. However, growing interest in
improving programoperations and expanding participation has led
USDA to commission a new study of the SFSP,known as the SFSP
Implementation Study. This report presents the results of the
SFSPImplementation Study, a descriptive study of the operations of
the SFSP at the state and locallevels. Mathematica Policy Research,
Inc. (MPR), under contract to the USDA EconomicResearch Service
(ERS), has collected nationally representative data to describe how
theprogram works and how SFSP staff feel it could be improved.
To help the reader understand the program and the issues
considered in the study, thischapter first provides background
information on the SFSP. It then describes the researchquestions,
the study design, and the data sources used in the rest of the
report.
A. THE SFSP
This section describes the SFSP and its history. It also
provides a brief overview of trendsin participation since the
1970s.
1. What Is the SFSP?
The SFSP was created to provide nutritious meals for children
from low-income families,particularly those who live in low-income
neighborhoods, when school is not in session. Theprogram operates
in all 50 states, the District of Columbia, the Virgin Islands, and
Puerto Rico.The USDA's Food and Nutrition Service (FNS) provides
federal funds to operate the SFSP, andstates distribute the funds
to local program sponsors, which are the organizations that carry
outthe day-to-day operations of the SFSP. Meals often are provided
in conjunction with educationaland/or recreational activities, such
as summer school, sports, arts and crafts activities, and camp.
a. Eligibility of Sponsors, Sites, and Children
Sponsor Eligibility. Organizations eligible to sponsor the SFSP
are public or privatenonprofit School Food Authorities
(SFAsgoverning bodies of schools or school districts thatoffer the
NSLP); public or private nonprofit residential summer camps; local,
municipal, county,
1
5
-
or state government units; public or private colleges or
universities that participate in theNational Youth Sports Program
(NYSP)'; and other private nonprofit organizations.
Privatenonprofit sponsors include youth organizations (such as
branches of the YMCA and Boys andGirls Clubs), religious
organizations, social service agencies, and other community
groups.
Site Eligibility. Each sponsor operates one or more SFSP sites
that serve meals to children.These sites may be school cafeterias,
parks, recreation centers, playgrounds, and other locations.Two
main types of sites operate under the SFSP: (1) "open" sites, and
(2) "enrolled" sites. Lesscommon types of sites are "camp" sites
and NYSP sites.
An open site must be located in a neighborhood in which at least
50 percent of the childrenlive in households that have incomes at
or below 185 percent of the federal poverty level. Mealsserved at
open sites must be provided to children at no charge.2 Sponsors
must use NSLPrecords from a local school (showing the percentage of
children certified for free or reduced-price school lunches) or
census block group records of household income to document
theeligibility of open sites for the SFSP.3
At enrolled sites, at least 50 percent of the children who are
enrolled at the site must live inhouseholds that are at or below
185 percent of poverty. However, sponsors are reimbursed formeals
served to all children who are enrolled. To demonstrate their
eligibility as enrolled sites,sponsors must collect either
enrollment forms documenting the family income of
participatingchildren or NSLP records from local schools
documenting the enrolled children's eligibility forfree or
reduced-price meals.
Residential summer camps may qualify as SFSP sites. Unlike
enrolled sites, however,residential camps are reimbursed only for
meals served to children from households withincomes at or below
185 percent of poverty, and they are not subject to the requirement
that atleast 50 percent of enrolled children be low-income.
Documentation required to show children'seligibility is the same as
for enrolled sites.
'The NYSP is a federally funded program that provides organized
instruction in athletics tolow-income children.
2A "restricted open" site is a new subcategory of open site.
Attendance at a restricted opensite initially is open to the broad
community but is later limited by the sponsor for security,safety,
or control reasons. For example, space limitations might require
that a restricted open siteserve only the first 100 children who
arrive each day. Children must be served on a
first-come,first-served basis at all open sites.
3According to current federal policy, sponsors should use NSLP
records, if possible. Aswith the SFSP, eligibility for the NSLP is
based on household income at or below 185 percent ofthe federal
poverty level. Therefore, an open site can demonstrate SFSP
eligibility if 50 percentof the children in the attendance area are
eligible to receive free or reduced-price NSLP meals,and, as
discussed in the text above, an enrolled site can demonstrate its
eligibility for the programif 50 percent of the children attending
the program are eligible for NSLP meals.
2
6
-
NYSP sites serve children enrolled in National Youth Sports
Programs. An NYSP site mayqualify for the program (1) as an
enrolled site, (2) by showing that attending children live inareas
that would qualify for open sites, or (3) by providing written
documentation that it meetsU.S. Department of Health and Human
Services (DHHS) income eligibility guidelines for theNYSP.4
Children's Eligibility. Children through age 18 are eligible to
receive SFSP meals.5Individuals older than age 18 may receive these
meals if they have physical or mental disabilitiesand if they
participate in special school programs for students with
disabilities.
b. Administrative Structure
The SFSP operates on three administrative levels: (1) the
federal FNS, (2) state agencies,and (3) local sponsors. Each level
has unique roles and responsibilities.
FNS. The FNS develops SFSP regulations and procedures and
distributes program fundsthrough its seven regional offices.
Regional FNS offices fund state SFSP programs on the basisof annual
state plans describing the size and scope of SFSP operations,
provide technicalassistance to states and sponsors, and monitor
state programs.
State Agencies. In most states, a state government agencyusually
the state educationagency that administers the school meals
programsadministers the SFSP. State agencies areresponsible for
recruiting sponsors, approving sponsors' applications, providing
training andtechnical assistance to sponsors and sites, monitoring
compliance with program rules andregulations, and handling claims
for reimbursement. Many of these activities are seasonal, as
theSFSP operates primarily during the summer.6 Many state agency
staff who work on SFSP workon other nutrition programs in the
nonsummer months.
When a state government has not assumed responsibility for
administering the program,FNS regional offices manage the program
and perform all the functions that state agencies wouldperform. In
2001, FNS regional offices administered the program only in
Michigan and Virginiaand for nonprofit and camp sponsors in New
York.
4DHHS is the primary grantor for the NYSP. To qualify for
participation in the NYSP underDHHS rules, at least 90 percent of
enrolled children must be at or below the federal povertylevel.
5A state agency must approve a sponsor's request to serve
infants. If infants are served, thesponsor must use the Child and
Adult Care Food Program's (CACFP's) meal requirements
forinfants.
6Some SFSP programs operate on a year-round basis. For example,
schools using a year-round schedule may use the SFSP to feed
children who are off-track (that is, who are on ascheduled school
break) throughout the year.
3
7
-
Sponsors. Program sponsors operate the SFSP at the local level
and have directresponsibility for either purchasing or preparing
meals and for monitoring program operationsand food safety and
quality. Sponsors must publicize their program; select meal sites;
hire andtrain staff; arrange for the preparation or delivery of
meals that meet the SFSP's standards;monitor meal attendance
levels; and maintain records and complete paperwork on
costs,attendance, food served, and related matters. Under federal
regulations, state agencies arerequired to make training available
to sponsors, and to ensure that sponsors receive the trainingthey
need.
Sponsors must submit SFSP applications to the state agency
annually. After an applicationhas been approved, the sponsor enters
into an agreement with the state to provide services.States monitor
sponsors' compliance with program rules; serious violations may
result in asponsor's termination or in the denial of its
application for the next program year. States mayalso disallow
meals for reimbursement purposes if the meals are not served
according to programrules.
Sponsors must train site supervisors on SFSP rules before their
sites open. Site staff mustserve meals to children in a way that
meets program requirements; supervise the children whilethey eat;
and accurately record the number of meals prepared or delivered,
the number served tochildren as firsts and as seconds, and the
number served to program staff. They also must followappropriate
procedures related to food safety and to storage or disposal of
leftovers. (Mealpattern requirements, which are intended to ensure
that SFSP meals provide good nutrition, arediscussed in detail in
Chapter V.)
c. Funding
The SFSP is funded through the USDA budget. Total federal
expenditures on the programwere $272 million in FY 2001 (Food and
Nutrition Service 2002a). The federal governmentfunds state
agencies for their costs of administering the program and also
funds sponsorreimbursements. States are not required to match
federal SFSP funds.
State agencies receive SFSP state administrative funds (SAFs) to
cover their administrativecosts. The SAF amount is calculated on
the basis of the total amount of program fundssponsors' operating,
sponsors' administrative, and health inspection fundsproperly
payable toa state for the SFSP in the preceding fiscal year (7CFR
225.5[a][1]). A state receivesadministrative funding equal to (1)
20 percent of the first $50,000 in program funds, (2)10 percent of
the next $100,000, (3) 5 percent of the next $250,000, and (4) 2.5
percent of anyadditional dollars in program funds that it received
during the previous fiscal year. FNS may
7Total federal expenditures include cash payments for meals
served, sponsors'administrative costs, the states' administrative
expenses, health inspection costs, and entitlementand bonus
commodity costs.
4
-
adjust the level of state administrative funding based on
changes in program size since theprevious year.8
Most sponsors are eligible to be reimbursed for the costs of
serving up to two meals orsnacks per child per day at their sites;
sites that are residential camps or sites serving migrantchildren
may receive reimbursement for up to three meals or snacks per child
per day. Sponsorreimbursements cover two types of costs: (1)
operational costs, including those associated withthe purchase,
preparation, and delivery or service of meals; and (2)
administrative costs,including office expenses, administrative
staff's salaries, program management costs, andinsurance. The
amount of reimbursement that sponsors receive is the lesser of (1)
actual netdocumented costs, or (2) the number of reimbursable meals
served multiplied by thecorresponding per meal reimbursement rates
for operating and administrative costs. In addition,administrative
cost reimbursements may not exceed the amount specified in the
budget approvedby the state agency. States may provide advance
funding to sponsors to help cover costs untiltheir reimbursement
claims are paid, but these advances are deducted from later
reimbursements.
The SFSP operating reimbursement rate in 2001 was $2.23 per
lunch or supper meal;breakfast and snacks were reimbursed at lower
rates (Table 1.1).9 Sponsors' maximum per-mealreimbursement rates
for administrative costs vary according to their location and/or
method ofmeal preparation (Table 1.1). SFSP meals may be prepared
by the sponsor or delivered by avendor; administrative
reimbursements are lower for meals purchased from a vendor and
servedat urban sites. For example, the 2001 SFSP administrative
reimbursement rate for a lunch orsupper was $0.2325 per meal for
rural sites and for urban self-preparation sites and was $0.1925per
meal for urban vended sites.
SFSP per-meal reimbursement rates (for administrative and
operating costs combined) arehigher than NSLP and SBP per-meal
reimbursement rates for students who qualify for free meals(which
include both operating costs and administrative costs in a single
rate). The size of thedifference depends on whether the SFSP
sponsor receives the higher or lower administrative
costreimbursement rate, and whether the NSLP or SBP rates are those
for an SFA in a highlydisadvantaged area. For lunch, the maximum
2001 SFSP combined reimbursement rate per meal($2.4625) was about
17 percent higher than the NSLP reimbursement rate for free lunches
for
8As discussed in detail in Chapter III, states commonly
supplement these funds with StateAdministrative Expense (SAE)
funds, a pool of federal funds that is used for state
agencyadministrative costs incurred in connection with other child
nutrition programs, including theNSLP, SBP, Special Milk Program,
and CACFP. These funds can be transferred to cover
SFSPadministrative costs that are not covered by SAF.
9Alaska's and Hawaii's rates are higher than those of the other
states and territories.
5
-
TABLE I.1
SFSP MAXIMUM PER-MEAL REIMBURSEMENT RATES FOR ALL STATESEXCEPT
ALASKA AND HAWAII, 2001
(Dollars)
Administrative Rates
Self-Preparation or OtherMeal Operating Rate Rural Sites
Sites
Breakfast 1.28 0.1275 0.1000
Lunch and Supper 2.23 0.2325 0.1925
Snack 0.52 0.0625 0.0500
SOURCE: Food and Nutrition Service (2002b).
NOTE: Rates effective January 1, 2001, through December 31,
2001.
-
school year 2001-2002 in heavily low-income areas ($2.11).1° The
higher reimbursement ratesfor the SFSP are one of several factors
that affect whether SFAs decide to operate the SFSP orthe NSLP
during the summer." In the case of other types of sponsors, the
higher rates reflect thefact that these sponsors are likely to have
higher costs than are SFAs, especially if they do notprovide food
service year-round.
2. History of the Program
The SFSP began in 1968 as a pilot program aimed at providing
meals to low-incomechildren during the summer. It was authorized as
a permanent program in 1975. Over time,changes in policy goals and
concerns have led to revisions in the eligibility criteria for
sponsorsand sites, administrative rules and regulations, and levels
of reimbursement. Participation hasfluctuated accordingly.
a. Late 1970s: Problems Led to Contraction
When it was authorized, the SFSP provided funding to open sites
located in areas in whichat least one-third of children came from
households with an income at or below 185 percent ofthe poverty
level. Subsequently, fraud and abuses in program administration
occurring duringthe mid-to-late 1970s were described in a series of
findings by the U.S. General AccountingOffice (GAO) and in the
media (U.S. General Accounting Office 1977 and 1978). Someprogram
operators were prosecuted. In addition to outright fraud, there
were reports of highlevels of food waste, poor-quality food, and
failure to meet meal pattern guidelines. Most abusesinvolved large
nonprofit sponsors. During the late 1970s, in response to concerns
about fraudand abuse, administrative oversight of sponsors
increased, sponsorship by nonprofitorganizations was limited, and
registration requirements for food service management companieswere
introduced (Food and Nutrition Service 2002c). The Omnibus Budget
Reconciliation Actof 1981 went further by prohibiting private
nonprofit groups (except private schools andresidential camps) from
serving as sponsors. In addition, for budgetary reasons, the
Act
1°The NSLP rate used applies to areas in which 60 percent or
more of meals are served freeor at a reduced-price. The maximum
SFSP breakfast reimbursement is about 3 percent higherthan the
maximum SBP reimbursements for "severe needs" districts ($1.4075
for SFSP ruralsponsors and self-preparation sponsors versus $1.37
for SBP severe needs districts). NSLP andSBP rates are from the
Federal Register (2001). All rates discussed are for the
contiguousUnited States.
11 Under current rules, SFAs may operate the NSLP during the
summer for children ineducational programs. They are reimbursed for
these meals at the free, reduced, or paid rates,according to the
children's income levels. If they choose to participate in the SFSP
instead, theymust open their sites to all children and are subject
to additional paperwork requirements, butthey can then receive the
higher reimbursement rate. The "Seamless Summer
Initiative"(discussed in Section I.2.d of this chapter) is testing
changes in these rules.
-
restricted eligible areas for open sites to those in which at
least 50 percent of children came fromhouseholds at or below 185
percent of the poverty level.
These changes led to a decline in participation in the SFSP
(Figure I.1). In 1976, before thechanges went into effect, average
daily attendance in July (the estimate of program participationused
by FNS) was more than 3 millionthe highest level of SFSP
participation ever reported.By 1982, participation had dropped to
fewer than 1.5 million children.
b. Late 1980s and Early 1990s: Renewed Interest in Expansion
Starting in the late 1980s, USDA worked with Congress and
advocacy groups to reverse thedecline in SFSP participation. At
around the same time, FNS undertook a national evaluation ofthe
SFSP (Oh Is et al. 1988). The evaluation suggested that the program
was operating withoutmajor problems. FNS increased outreach and
technical assistance to potential sponsors andcollaboration with
nutrition advocates. In addition, a major nutrition policy advocacy
group, theFood Research and Action Center (FRAC), undertook a
campaign to expand the SFSP in 1991that has continued to the
present. FRAC has published an annual report on SFSP
participationsince 1993, titled "Hunger Doesn't Take a Vacation"
(Food Research and Action Center 2002).
Several pieces of legislation were passed to address the
reduction in participation. In 1989,private nonprofit sponsors were
allowed back into the program, subject to provisions forexpanded
state agency training and monitoring and with restrictions on their
operations. A majorobjective of this action was to increase program
coverage (Decker et al. 1993). Two studiesconducted by MPR at this
time found that the reentry of private nonprofit sponsors
contributedsignificantly to an increase in the number of sponsors
and sites between 1989 and 1991, as wellas to an increase in
program attendance and in the number of SFSP meals served (Decker
et al.1990; and Decker et al. 1993).
In 1994, special grants were established for sponsors' start-up
and expansion costs, whichprovided funds in addition to meal
reimbursements. Some of the restrictions on private
nonprofitsponsors were relaxed as well. For example, nonprofit
sponsors operating in areas formerlyserved by school or government
sites no longer were subject to a 1-year waiting period.
c. The Late 1990s: Reimbursement Cuts and Streamlined
Administration
In 1996, the Personal Responsibility and Work Opportunity
Reconciliation Act (PRWORA)mandated major changes to many social
welfare programs, including cutbacks in the Food StampProgram and
in several child nutrition programs, one of which was the SFSP.
PRWORAremoved expansion of the SFSP as a stated program goal,
reduced reimbursement rates and thenumber of reimbursable meals
allowed per day, and eliminated start-up and expansion grants
forsponsors. In a GAO study conducted shortly after the changes
took effect, some sponsorsreported that they had substituted less
expensive foods for those previously served, reduced staffwages,
and reduced the number of sites they operated (U.S. General
Accounting Office 1998).The study also reported that, overall, the
reductions in reimbursements had little effect on thenumber and
characteristics of sponsors participating in the program or on the
number of
-
3,500,000
3,000,000
2,500,000
2,000,000
1,500,000
1,000,000
500,000
0
FIGURE I.I
SFSP PARTICIPATION AND PROGRAM MILESTONES
Reimbursement rates andnumber of reimbursablemeals per day
lowered; start-upgrants ended (1996)
Grants established forsponsors' start-up andexpansion costs
(1994)
Nonprofit sponsorsprohibited; site eligibilityrestricted
(1981)
Problems led toincreasedsponsoroversight;nonprofitsponsors
limited(late 1970s)
rr
Nonprofi sponsorsreadmitted (1989)
NO. woo =V
Most remaining restrictionson nonprofit sponsors
removed;applications and administrativereviews for
experiencedsponsors streamlined(late 1990s)
0
Summer participationin the NSLP
T I I
kr, N rn ., kr, 1.4 le -1 1.4N N N 00 00 00 00 00 (7,rn
goo!(7, rn a e7,
Year
SOURCE: SFSP participation (Food and Nutrition Service 2002a);
NSLP July participation is fromunpublished data provided by the FNS
Child Nutrition Division, May 7, 2002.
-
participants. However, the GAO study examined the SFSP only
during the first year after thechanges took effect.
During the late 1990s, most of the remaining restrictions on
private nonprofit sponsors wereremoved.' Changes were made to
streamline application procedures for sponsors thatsuccessfully had
participated in the SFSP during the prior year. Other changes
focused statemonitoring requirements on new sponsors, large
sponsors, sponsors that had operationalproblems, and sponsors
experiencing frequent staff turnover. The federal requirement
forregistering food service management companies was removed,
thereby giving states discretion asto whether they continue this
practice (Food and Nutrition Service 1998).
d. Recent Federal and State Initiatives
Currently, states and FNS have renewed their focus on how to
increase the availability ofsummer meals to children in low-income
areas (Food and Nutrition Service 2002d). Both FNSand the states
are experimenting with new approaches to expanding the program.
At the federal level, two initiatives began in summer 2001: (1)
a pilot project, mandated byCongress, to simplify reimbursement
rates in certain states; and (2) the Seamless SummerFeeding
Waivers, an initiative designed to encourage school districts to
serve more meals duringthe summer In late 2000, Congress mandated a
3-year pilot project as part of the Richard B.Russell School Lunch
Act (Public Law 106-554). The "14-state" pilot project, which began
insummer 2001 (as this study went into the field), applies to
sponsors in 13 states and Puerto Rico,jurisdictions which have low
SFSP participation levels. One of the goals of the pilot is to test
anapproach to reducing paperwork for sponsors (Food and Nutrition
Service 20020. The pilotprovides the maximum reimbursement rate to
sponsors in these jurisdictions. It also removes thedivision
between administrative and operating cost reimbursements; sponsors
receive onereimbursement to cover all of their expenses, whether
administrative or operating. Sponsors nolonger have to track
administrative and operating costs separately, or to report costs
to the stateagency. They earn "meals times rates," which makes it
easier for them to estimate the amount offunding they will receive.
The pilot applies to most sponsors; however, it excludes
privatenonprofit sponsors that are not schools or residential
camps.
Another federal initiativethe Seamless Summer Food Waiverseeks
to encourage moreschool districts to serve more meals to children
in low-income areas during the summer. Underthis waiver, school
districts use NSLP meal service and claims procedures to provide
summermeal service to children aged 18 or younger at sites that
meet SFSP criteria. School districtsoperating these waiver sites
are subject to the less complicated administrative requirements
ofthe NSLP. All meals served at the waiver sites are claimed as
NSLP meals and are reimbursed atthe NSLP free rate, including the
allowance for commodities. Sponsors do not receive a separate
12Nonprofit SFSP sponsors continue to be limited to operating no
more than 25 sites.However, the National School Lunch Act provides
authority for FNS to approve waivers fromthis limit.
10
14
-
administrative reimbursement for these meals. The seamless
summer initiative began in 2001with initial waivers granted to five
school districts (two in California and three in Florida).Starting
in 2002, school districts nationwide may request a waiver through
their state educationagency to operate a seamless summer feeding
program through summer 2004 (U.S. Departmentof Agriculture
2002a).
Several states have implemented laws to require school districts
in low-income areas tooperate the SFSP; because these initiatives
are relatively recent, however, little information isavailable on
how they are working. Texas state law requires public school
districts in which atleast 60 percent of children are eligible for
free or reduced-price school meals to sponsor theSFSP or to arrange
for a sponsor in their enrollment areas. There is no minimum,
however, onthe number of days or the number of sites at which the
program must operate. Furthermore,school districts may receive a
waiver exempting them from this responsibility if they
candemonstrate that the costs of running the program will exceed
available funds, that renovationsor construction in the district
prevent them from operating the program, or that other
extenuatingcircumstances exist. Texas also provides funds for
outreach and for supplemental mealreimbursements. Missouri also has
a mandate to operate the SFSP in high poverty areas of thestate
(Food Research and Action Center 2002).
Six states in addition to Texas provided state funding for the
SFSP in 2001 (Food Researchand Action Center 2002).13 Some states
provided supplemental reimbursements for all sponsors,some provided
start-up and expansion grants, and some provided funding for
outreach.
3. Participation and Participation Rates
At the time of its authorization as a permanent program in 1975,
the SFSP served almost1.8 million children per day during the peak
summer month of July (Figure I.1).14 In 1976,before the tightening
of program regulations and of restrictions on nonprofit
sponsors,participation rose to almost 3.5 million. By 1982, it had
fallen below 1.5 million. Starting in1983, participation rose
slowly but steadily, reaching 2 million children in 1993. Since
then,despite declining slightly since 1999, participation has been
relatively stable, hovering between2.1 and 2.2 million
children.
"The states are California, Massachusetts, Minnesota, New York,
Vermont, andWashington.
14FNS measures participation as average daily attendance
reported by sponsors operating inJuly, the peak month for SFSP
participation. State agencies collect these data and report them
toFNS. Because the program does not track individual children who
participate, it is not possibleto determine the number of children
who ever receive meals during the summer.
11
15
-
CONTENTS
Chapter Page
I INTRODUCTION 1
A. THE SFSP 1
1. What Is the SFSP? 12. History of the Program 73.
Participation and Participation Rates 11
B. STUDY OBJECTIVES AND RESEARCH QUESTIONS 12
C. STUDY DESIGN AND DATA SOURCES 15
1. Study Design 152. Data Collection 16
D. PLAN OF THE REPORT 18
II PROGRAM CHARACTERISTICS 19
A. CHARACTERISTICS OF STATE AGENCIES 20
B. CHARACTERISTICS OF SPONSORS 23
1. Overview of Sponsors' Characteristics 232. Characteristics of
Different Types of Sponsors 293. Changes in Characteristics of
Sponsors Since 1986 32
C. CHARACTERISTICS OF SITES 38
1. Overview of Sites 382. Comparison of Sites with School
Sponsors and Sites with Nonschool
Sponsors 433. Changes in Site Characteristics Since 1986 45
D. DEMOGRAPHIC CHARACTERISTICS OF PARTICIPANTS 48
E. SITE SCHEDULING AND TRANSPORTATION ISSUES 51
16
-
CONTENTS (continued)
Chapter Page
III PROGRAM ADMINISTRATION 57
A. STAFFING AND FUNDING 58
1. State-Level Staffing and Funding 592. Sponsor Staffing and
Funding 633. State Payments to Sponsors 73
B. SPONSOR APPLICATIONS 77
C. TRAINING AND TECHNICAL ASSISTANCE 82
1. State Training of Sponsors 822. State Technical Assistance to
Sponsors 853. Sponsor Training and Technical Assistance to Sites
91
D. PROGRAM MONITORING 94
1. State Monitoring of Sponsors and Sites 972. Sponsors'
Monitoring of Sites 101
E. VENDOR/FOOD SERVICE MANAGEMENT 104
1. Use of Vendors 1042. Reasons Why Most Sponsors Prepared Meals
1053. Reasons Why Some Sponsors Contracted with Vendors 1054.
Selecting and Monitoring Vendors 109
IV PROGRAM PARTICIPATION AND OUTREACH 113
A. STAFFS' VIEWS ON PARTICIPATION LEVELS 114
1. State Administrators' Views on Participation 1142. Sponsors'
and Former Sponsors' Views on Participation and Program
Expansion 1173. Site Supervisors' Views on Capacity Constraints
and Barriers to
Participation 123
vi
17
-
CONTENTS (continued)
Chapter Page
B. EFFORTS TO EXPAND THE PROGRAM 126
1. State Agency Outreach 1262. Sponsors' Efforts to Increase
Participation at Their Sites 1313. Sites' Efforts to Increase
Participation 137
C. NEW SPONSORS 137
D. SPONSORS THAT LEAVE THE PROGRAM 141
1. State Agencies' Strategies to Promote Sponsor Retention 1412.
Characteristics of Former Sponsors and Current Sponsors 1433.
Reasons Why Former Sponsors Left the Program 146
V MEAL SERVICE 157
A. CHARACTERISTICS OF MEAL SERVICE 159
1. Meal Service Characteristics 1592. Disposition of Available
Meals 1613. Handling and Storage of Food 1644. Meal Order
Adjustment and Transport of Food Prepared Off Site 166
B. CONTENT OF MEALS SERVED 168
1. Most Frequently Served Foods 1682. Food Preferences of
Participants 1763. Nonschool Sponsors' Compliance with SFSP Meal
Pattern 1794. Nutrient Content of Meals Served 186
C. EXTENT OF PLATE WASTE 202
1. Nutrients Wasted on Plates 2022. Foods Wasted on Plates 2103.
Reasons for Food Waste 215
vii
18
-
CONTENTS (continued)
Chapter Page
VI SUMMARY AND CONCLUSIONS 217
A. PROGRAM ADMINISTRATION 217
1. Key Findings 2182. Issues for the Future 219
B. PARTICIPATION 220
1. Key Findings 2202. Issues for the Future 2213. Future
Research 222
C. MEAL SERVICE 223
1. Key Findings 2232. Issues for the Future 2243. Future
Research 227
REFERENCES 229
APPENDIX A: DATA COLLECTION METHODS A.1
APPENDIX B: SAMPLING AND SAMPLE WEIGHTS B.1
APPENDIX C: PROFILES OF SFSP SITES C.1
APPENDIX D: DETAILED TABULATIONS FROM THE 2001SPONSOR-SITE
DATABASE D.1
APPENDIX E: NUTRIENT AND FOOD CODING ANALYSIS E. 1
APPENDIX F: SUPPLEMENTAL MEAL SERVICE TABLES F.1
APPENDIX G: SUPPLEMENTAL TABLES FOR CHAPTER III G.1
viii
19
-
TABLES
Table
I.1 SFSP MAXIMUM PER-MEAL REIMBURSEMENT RATES FOR ALL
Page
STATES EXCEPT ALASKA AND HAWAII, 2001 6
1.2 DATA COLLECTION FOR THE SFSP IMPLEMENTATION STUDY 17
II.1 CHARACTERISTICS OF STATE AGENCIES 21
11.2 SELECTED CHARACTERISTICS OF SPONSORS 24
11.3 SELECTED CHARACTERISTICS OF SPONSORS, BY TYPE OFSPONSOR
30
11.4 SELECTED CHARACTERISTICS OF SPONSORS, BYSCHOOL/NONSCHOOL
SPONSOR 33
11.5 NUMBER OF SFSP SPONSORS, SITES, AND PARTICIPANTS, BYTYPE OF
SPONSOR, JULY 1986 AND JULY 2001 35
11.6 CHANGES IN SELECTED CHARACTERISTICS OF SPONSORS SINCE1986
37
11.7 SELECTED SITE CHARACTERISTICS 39
11.8 SELECTED CHARACTERISTICS OF SITES, BYSCHOOL/NONSCHOOL
SPONSOR 44
11.9 CHANGES IN SELECTED SITE CHARACTERISTICS SINCE 1986 46
II.10 DEMOGRAPHIC CHARACTERISTICS OF PARTICIPANTS 49
II.11 DEMOGRAPHIC CHARACTERISTICS OF PARTICIPANTS IN 1986AND
2001 50
11.12 SITE SCHEDULE AND ATTENDANCE 52
11.13 FACTORS AFFECTING DAY-TO-DAY VARIATION INPARTICIPATION AT
SITE 53
11.14 TRANSPORTATION TO SITE 54
11.15 CHARACTERISTICS OF SITES THAT PROVIDE TRANSPORTATION
56
III.1 STATE-LEVEL STAFFING 60
ix20
-
TABLES (continued)
Table Page
111.2 STATE AGENCIES' REPORTS ON THE ADEQUACY OF THEIRSTAFFING,
BY FUNCTION AND OVERALL 61
111.3 CHANGES IN STATE AGENCY STAFFING 62
111.4 STATE AGENCY FUNDING 64
111.5 SPONSORS' REPORTS ON THE ADEQUACY OF THEIR STAFFING,BY
FUNCTION AND OVERALL 66
111.6 SPONSORS' REPORTS ON SFSP REIMBURSEMENTS AND OTHERFUNDING
SOURCES 69
111.7 EXPERIENCED SPONSORS' COST-CONTROL STRATEGIES 72
111.8 STATE AGENCIES' VIEWS ON THE USE OF ADVANCE ANDSTART-UP
FUNDS 75
111.9 STATE ADMINISTRATORS' VIEWS ON THE PILOT PROJECT 76
III.1 0 PROCESSING OF SPONSOR APPLICATIONS 80
III.1 1 SPONSORS' COMMENTS ON THE APPLICATION PROCESS 81
111.12 STATE TRAINING OF SPONSORS 83
111.1 3 NUMBER OF SPONSORS' STAFF ATTENDING STATE TRAINING
84
111.14 TOPICS COVERED IN STATE TRAINING OF SPONSORS 86
111.1 5 STATE ADMINISTRATORS' PERSPECTIVE ON TOPICS
GIVINGSPONSORS DIFFICULTY DURING OR AFTER STATE TRAINING,BY
SPONSORS' EXPERIENCE 88
111.1 6 AREAS OF STATE AGENCIES' TECHNICAL ASSISTANCE 89
111.17 SPONSORS' VIEWS ON TECHNICAL ASSISTANCE FROM
STATEAGENCIES 90
111.1 8 SPONSOR-PROVIDED TRAINING FOR SITE STAFF 92
111.19 SITE SUPERVISORS' VIEWS ON TRAINING AND
TECHNICALASSISTANCE FROM SPONSORS 95
111.20 STATE AGENCIES' REVIEWS OF SPONSORS AND SITES 98
x
21
-
TABLES (continued)
Table Page
111.21 MULTISITE SPONSORS' MONITORING OF SITES 102
111.22 SPONSOR TYPE, BY SELF-PREPARED AND VENDED MEALS 106
111.23 REASONS SPONSORS PREPARE MEALS RATHER THANCONTRACT WITH A
VENDOR 107
111.24 SPONSORS' VIEWS ON ADVANTAGES AND DISADVANTAGES OFVENDORS
108
111.25 VENDOR BIDS AND VENDOR MONITORING 110
IV.1 STATE ADMINISTRATORS' VIEWS ON WHY PARTICIPATION ISLOW
116
W.2 SPONSORS' IDENTIFICATION OF THE MAIN BARRIERS TOINCREASED
PARTICIPATION 118
W.3 SPONSORS' INTEREST IN EXPANDING THE NUMBER OF SITES 119
IV.4 REASONS FOR SPONSORS' LACK OF INTEREST IN EXPANDINGTHE
NUMBER OF SITES 121
W.5 SPONSORS' VIEWS ON INCREASING THE LENGTH OF THESUMMER
SESSION 122
W.6 SITE CAPACITY CONSTRAINTS ON PARTICIPATION 124
IV.7 SITE SUPERVISORS' VIEWS ON BARRIERS TO
CHILDRENS'PARTICIPATION AT SITE 125
IV.8 STATE AGENCIES' WORK WITH PARTNER ORGANIZATIONS 127
W.9 STATE AGENCIES' EFFORTS TO ATTRACT SPONSORS 129
W.10 SPONSORS' STRATEGIES TO INCREASE PARTICIPATION 132
W.11 FORMER SPONSORS' EFFORTS TO INCREASE PARTICIPATION 134
IV.12 SPONSORS' WORK WITH OTHER ORGANIZATIONS TO PROMOTEAND
SUPPORT THE SFSP 135
W.13 SITE SUPERVISORS' VIEWS ON PUBLICITY EFFORTS 138
xi
22
-
TABLES (continued)
Table Page
IV.14 COMPARISON OF SELECTED CHARACTERISTICS OF NEW
ANDCONTINUING SPONSORS 139
IV.15 STATE AGENCIES' EFFORTS TO RETAIN SPONSORS 142
IV.16 COMPARISON OF SELECTED CHARACTERISTICS OF FORMERSPONSORS
AND CURRENT SPONSORS 144
IV.17 STATE ADMINISTRATORS' VIEWS ON WHY SPONSORS LEFT
THEPROGRAM 147
IV.18 REASONS FORMER SPONSORS CITED FOR LEAVING THEPROGRAM
150
IV.19 CHANGES THAT COULD INDUCE FORMER SPONSORS TORETURN TO THE
PROGRAM 153
IV.20 FORMER SPONSOR SITES PICKED UP BY ANOTHER SPONSOR 156
V.1 SELECTED MEAL CHARACTERISTICS 160
V.2 DISPOSITION OF AVAILABLE MEALS 162
V.3 HANDLING AND STORAGE OF FOOD 165
V.4 MEAL ORDER ADJUSTMENT AND TRANSPORT OF FOODPREPARED OFF SITE
167
V.5 FOODS MOST COMMONLY SERVED AT BREAKFAST, BY FOODCATEGORY
169
V.6 FOODS MOST COMMONLY SERVED AT LUNCH, BY FOODCATEGORY 171
V.7 FOOD PREFERENCES OF PARTICIPANTS, REPORTED BY
SITESUPERVISORS 177
V.8 MEAL COMPLIANCE AT BREAKFAST FOR NONSCHOOLSPONSORS 182
V.9 MEAL COMPLIANCE AT LUNCH FOR NONSCHOOL SPONSORS 183
V.10 MEAN ENERGY AND KEY NUTRIENTS SERVED AT SFSPBREAKFASTS AND
COMPARISON WITH RDAs 189
xii
23
-
TABLES (continued)
Table Page
V.11 MEAN AND DISTRIBUTION OF ENERGY AND KEY NUTRIENTSSERVED ON
A SINGLE DAY AT BREAKFAST 190
V.12 MEAN ENERGY AND KEY NUTRIENTS SERVED AT SFSPLUNCHES AND
COMPARISON WITH RDAs 192
V.13 MEAN AND DISTRIBUTION OF ENERGY AND KEY NUTRIENTSSERVED ON
A SINGLE DAY AT LUNCH 194
V.14 MEAN ENERGY AND KEY NUTRIENTS SERVED AT SFSP SUPPERSAND
COMPARISON WITH RDAs 195
V.15 MEAN AND DISTRIBUTION OF ENERGY AND KEY NUTRIENTSSERVED ON
A SINGLE DAY AT SUPPER 197
V.16 MEANS FOR OTHER NUTRIENTS SERVED AT SFSP BREAKFASTSAND
COMPARISON WITH RDAs 198
V.17 MEANS FOR OTHER NUTRIENTS SERVED AT SFSP LUNCHES
ANDCOMPARISON WITH RDAs 199
V.18 MEANS FOR OTHER NUTRIENTS SERVED AT SFSP SUPPERS
ANDCOMPARISON WITH RDAs 200
V.19 COMPARISON OF MEAN NUTRIENT PROFILES FOR SFSP MEALSAND
SCHOOL MEALS 201
V.20 MEAN AND PERCENTAGE OF NUTRIENTS WASTED ATBREAKFAST, BASED
ON PLATE WASTE OBSERVATIONS 203
V.21 MEAN AND PERCENTAGE OF NUTRIENTS WASTED AT LUNCH,BASED ON
PLATE WASTE OBSERVATIONS 205
V.22 MEAN AND PERCENTAGE OF NUTRIENTS WASTED AT SUPPER,BASED ON
PLATE WASTE OBSERVATIONS 207
V.23 PERCENTAGE OF MOST COMMONLY SERVED BREAKFASTFOODS WASTED,
BY FOOD CATEGORY 211
V.24 PERCENTAGE OF MOST COMMONLY SERVED LUNCH FOODSWASTED, BY
FOOD CATEGORY 213
24
-
FIGURES
Figure Page
I.1 SFSP PARTICIPATION AND PROGRAM MILESTONES 9
1.2 SFSP PARTICIPATION AND NSLP FREE OR
REDUCED-PRICEPARTICIPATION, 1975 THROUGH 2001 13
1.3 SFSP PARTICIPATION RATE 14
xv
25
-
Like the SFSP, the NSLP provides meals to children during the
summer.I5 NSLPparticipants are children attending summer school and
on-track students at year-round schools.In line with national
trends toward year-round schooling and increased requirements for
summerschool participation, summer participation in the NSLP has
more than doubled during the pastdecade (see Figure I.1). Recently,
NSLP participation rose from 1.06 to 1.14 million betweenJuly 1999
and July 2000, more than offsetting a small decrease in the number
of children fedthrough the SFSP. Thus, recent decreases in SFSP
participation do not necessarily reflect fewerchildren receiving
summer meals through federal nutrition programs overall. Rather,
they mayreflect a redistribution in the number of children
accessing specific meal programs.16
The difference between the number of children who participate in
the SFSP and the numberwho participate in the free or reduced-price
component of the NSLP during the school yearalways has been large.
For example, in 1999, an average of 15 million children from
low-income households received free or reduced-price NSLP lunches
each month during the schoolyear, whereas only 2.2 million received
meals through the SFSP in July (Figure 1.2). The ratio ofSFSP
participants to free or reduced-price NSLP participants, converted
to a percentage, can beinterpreted as an approximate participation
rate in the SFSP among low-income children. Thispercentage was
relatively stable between 1989 and 1996, varying only from 14.5
percent to15.8 percent (Figure 1.3). It has been declining slightly
but continually since then. In 2000, theparticipation rate was less
than 14 percent for the first time since 1989. These recent
decreasesreflect both the slight decline in SFSP participation and
an increase in school-year participationin the NSLP by low-income
children.
B. STUDY OBJECTIVES AND RESEARCH QUESTIONS
The main objective of this study is to describe the operations
of the SFSP at the state,sponsor, and site levels, using nationally
representative data. The specific research questionsinclude:
1. How does the SFSP operate at the state, sponsor, and site
levels? Is the programoperating as intended by current policy and
regulations? What areas do staff believeare in need of
improvement?
15The CACFP is another year-round nutrition program that serves
meals to children in thesummer. Average daily participation in the
CACFP was approximately 2.65 million in 2001(including children and
adults served in day care homes and in child care centers).
According todata provided by the FNS Child Nutrition Division,
participation levels in the summer are similarto those throughout
the rest of the year, suggesting that the CACFP does not generally
serve theschool-aged children served by the NSLP during the school
year.
16As the Seamless Summer Food Waivers are implemented more
widely, this redistributionmay increase, which implies that it will
be increasingly important to examine NSLP and SFSPparticipation
jointly.
-
16,0
00,0
00
14,0
00,0
00
12,0
00,0
00
g 10
,000
,000
8,00
0,00
0
6,00
0,00
0
4,00
0,00
0
2,00
0,00
0 0
FIG
UR
E 1
.2
SFSP
PA
RT
ICIP
AT
ION
AN
D N
SLP
FRE
E O
R R
ED
UC
ED
-PR
ICE
PAR
TIC
IPA
TIO
N, 1
975
TH
RO
UG
H 2
001
og
mm
ll...
..
,,,0,
0
0
I'or
rm. m
ar m
m""
'' "
lial .
..,-
. Ina
....
.,.G
o.
=
(1
ii
1I
I1
II
fI
I
1975
1980
1985
1990
1995
SOU
RC
E: S
FSP
prog
ram
dat
a (F
ood
and
Nut
ritio
n Se
rvic
e 20
02a)
; NSL
P pr
ogra
m d
ata
(Foo
d an
d N
utri
tion
Serv
ice
2002
e).
2000
NSL
P
SFSP
NoT
E:
Dat
a re
flec
t 9-m
onth
ave
rage
s fo
r fr
ee o
r re
duce
d-pr
ice
part
icip
atio
n in
the
NSL
P (s
umm
er m
onth
s ar
e ex
clud
ed)
and
July
par
ticip
atio
n in
the
SFSP
. Dat
a fo
r 20
01 a
re p
relim
inar
y.
-
35 30 25 20a) on a
15 10 5 0
FIG
UR
E 1
.3
SFSP
PA
RT
ICIP
AT
ION
RA
TE
(Num
ber
of S
FSP
Part
icip
ants
as
a Pe
rcen
tage
of
Free
or
Red
uced
-Pri
ce N
SLP
Part
icip
ants
)
1975
1980
1985
1990
1995
2000
Yea
r
SOU
RC
E: S
FSP
prog
ram
dat
a (F
ood
and
Nut
ritio
n Se
rvic
e 20
02a)
; NSL
P pr
ogra
m d
ata
(Foo
d an
d N
utri
tion
Serv
ice
2002
e).
NO
TE
S:D
ata
refl
ect J
uly
part
icip
atio
n in
the
SFSP
div
ided
by
9-m
onth
ave
rage
s fo
r fr
ee o
r re
duce
d-pr
ice
part
icip
atio
n in
the
NSL
P (s
umm
er m
onth
s ar
eex
clud
ed).
Dat
a fo
r 20
01 a
re p
relim
inar
y.
-
2. What factors affect participation by sponsors and children?
What do program staffsee as the most important barriers to
participation? What efforts are they making toexpand participation?
What are the levels of entry and exit of program sponsors?Why do
some sponsors leave the program, and how do their characteristics
comparewith those of sponsors overall?
3. What is the nutritional quality of meals served, and what is
the extent of platewaste? How are SFSP meals prepared and served?
What are the foods served andportion sizes? How does the
nutritional content of SFSP meals compare with relevantnutrition
standards? What factors are associated with more nutritious meals
and lesswaste?
C. STUDY DESIGN AND DATA SOURCES
This section provides an overview of the design of the SFSP
Implementation Study, the datasources used, and the methods used to
collect the data. Appendix A and Appendix B discuss,respectively,
data collection in detail and the sample design and weights used in
the study.
1. Study Design
ERS worked with MPR to design this study under a previous
contract (Briefel et al. 2000).Key requirements of the design
included nationally representative data on sponsors,
formersponsors, and sites and the ability to link sponsor and site
data. The final study design includedthe following components:
Telephone interviews with all SFSP state administrators
A mixed-mode mail/telephone survey of a national sample of
sponsors
A telephone survey of former sponsors, defined as organizations
that sponsored theSFSP in 2000 but not in 2001
Visits by interviewer-observers to a national sample of program
sites sponsored bythe programs in the sponsor sample, which
included in-person interviews with sitesupervisors, structured
observations of program operations, and coding of a randomsample of
meals served and plate waste
The design for this study was influenced heavily by the last
comprehensive nationalevaluation of the SFSP (also conducted by
MPR), which collected data in 1986 (Ohls et al.1988). The study by
Ohls et al. collected data at the state, sponsor, and site levels
anddemonstrated that on-site observers could collect data on the
content of meals served and platewaste. Whenever appropriate, the
results of the current study are compared with the 1986 data.
In designing the current study, ERS asked MPR to consider the
feasibility of studyingfactors associated with participation. The
designs developed included (1) geocoding of thelocations of SFSP
sponsors and sites and geographic analysis of the characteristics
of areas
15
-
served and not served, and (2) a survey of families in areas
served by SFSP sites that wouldinclude both participants and
nonparticipants (Gordon et al. 2000). ERS decided not to pursuethe
participant-nonparticipant survey because of concerns about
feasibility. ERS is planning afuture study that will include the
geographic analysis of data on sponsor and site addresses
inrelation to census poverty data, which will allow USDA to assess
how well the SFSP is reachingareas in which low-income children
live.
2. Data Collection
Data collection for this study took place during the spring,
summer, and fall of 2001. Thedata collection for the study
consisted of four surveysa state administrator survey, a
sponsorsurvey, a former sponsor survey, and a site supervisor
surveyplus in-person observations ofoperations at a national sample
of sites and a large effort to compile administrative data.
Developing the sample frames for the study was a major
undertaking. State SFSP agenciesprovided lists of year 2000
sponsors in spring 2001 and subsequently provided lists of new
2001sponsors and lists of sponsors that had left the program. These
lists comprised the sample framesfrom which the samples of sponsors
and of former sponsors were selected. After sponsors wereselected
for the sample, they were contacted and asked to provide lists of
their sites; the listsbecame the sample frame for selecting the
sample of sites to visit. The level of cooperationreceived at all
levels was very high.
Table 1.2 describes the four surveys and the site observations.
The table shows their modeof data collection, sample sizes, and
response rates. To reduce respondent burden, all foursurveys asked
some questions about administrators' problems or challenges as
open-endedquestions. However, because respondents generally mention
only a few issues in response toopen-ended questionsmost often, the
ones about which they feel stronglyresponses to suchquestions
likely provide lower bounds on the numbers of administrators who
believe particularchallenges or problems are important. Throughout
this report, tables indicate when data reflectresponses to
open-ended questions.
During the fall and winter of 2001, SFSP state agencies also
provided MPR with detailedlists of all their 2001 SFSP sponsors and
lists of each sponsor's sites. These lists includedaddresses of
sponsors and sites that have been geocoded for future analysis, as
well asdescriptions of basic program characteristics.° The data
collected on sponsors included the typeof sponsor (school,
government, residential camp, NYSP, other nonprofit), dates of
operation,types of meals served, whether the sponsor was new to the
program, and meal counts for eachmeal. The data collected on sites
included dates of operation, types of meals served, an estimateof
average daily attendance for each meal, and an indicator of whether
the site was open orenrolled. These data, which represent a census
of the SFSP as of summer 2001, have beencompiled into a linked
database, the "SFSP 2001 Sponsor-Site Database."
"Decision Demographics, Inc., under subcontract to MPR, geocoded
the addresses in thedatabase.
16
30
-
TABLE 1.2
DATA COLLECTION FOR THE SFSP IMPLEMENTATION STUDY
Survey Mode of Data CollectionSample Response Rate
Size (Percentage)
State AdministratorCensus
Sponsor Survey
Former SponsorSurvey"
Site SupervisorSurvey and SiteObservations
Telephone interview 54a 100
Mail survey with telephone followup 126
Telephone interview 131
96
89
In-person interview and structured 162 95observations of site
operations, participantcharacteristics, content of meals served,and
plate waste
NOTE: See Appendix A for additional details on data
collection.
aThere are 54 state agencies to represent all 50 states, Puerto
Rico, Virgin Islands, and theDistrict of Columbia. The program in
New York state is partly run by the State Department ofEducation
and partly run by the FNS regional office, so both agencies were
included in thesurvey.
"Former sponsors are defined as agencies that were SFSP sponsors
in 2000 but not in 2001.
17
31
-
D. PLAN OF THE REPORT
Chapter II describes the SFSP at the state, sponsor, and site
levels and its participants, basedon data from the Sponsor-Site
Database and the various surveys. It also considers how
thecharacteristics of the program have changed since 1986, when
data for the previous study werecollected.
Each of the next three chapters addresses one of the three
research questions in turn.Chapter III describes how the program
was administered at the state, sponsor, and site levels, aswell as
areas of program operations that staff believed could be improved.
Chapter IV discussesparticipation and outreach, including staffs'
views on barriers to participation, the extent ofoutreach efforts,
and the reasons why sponsors left the program. Chapter V describes
the mealsserved in the SFSP, their nutrient content, and the extent
of plate waste. Finally, Chapter VIsummarizes the findings and
considers issues for the future.
18
32
-
II. PROGRAM CHARACTERISTICS
This chapter presents a statistical picture of the Summer Food
Service Program (SFSP) in2001 that covers the characteristics of
the state agencies, sponsors, and sites that operated theprogram.'
The chapter also describes SFSP participants and discusses factors
that may affecttheir participation. Many of the discussions compare
the SFSP in 2001 with the program as itwas 15 years earlier, when
the last comprehensive study was undertaken. The highlights of
thefindings are:
Half the program was school-sponsored in 2001. School Food
Authorities (SFAsthe governing bodies of schools or school
districts that operate the National SchoolLunch Program [NSLP])
comprised roughly half of all sponsors, ran about half of allsites,
and served about half of all meals. The number of school sponsors
operating inJuly almost tripled from 1986 to 2001, and their
average daily attendance increasedby 66 percent.
Government agencies (usually municipal recreation or social
services departments)constituted 14 percent of sponsors, but they
were the largest sponsors, on average,and served 31 percent of
meals. Residential camp sponsors were about 16 percent ofall
sponsors but served only 7 percent of meals. Because the number of
governmentand camp sponsors and the number of meals they served had
changed little from 1986to 2001, these sponsor types are,
proportionately, smaller parts of the program than in1986.
Nonprofit organizations, which have rejoined the program since
1986, represented18 percent of all sponsors in 2001. However, they
generally are restricted in size tono more than 25 sites and served
just 10 percent of all meals.
Sponsors that obtained meals from vendors comprised 18 percent
of sponsors, butthey operated 36 percent of sites and served 30
percent of meals. SFA vendorsprovided about one-third of vended
meals, and private vendors providedabout two-thirds.
Since 1986, the number of sites that provide breakfast and the
number that stay openfor longer than 6 weeks have increased. Almost
all sites (93 percent) offeredactivities in addition to meals, and
most (61 percent) were open for longer than6 weeks. About half of
all sites served breakfast, and essentially all served lunch.
'Appendix C provides a qualitative description of the SFSP,
through in-depth profiles ofnine sites selected to convey the wide
variability in the program. Readers who are not familiarwith the
program may find that this appendix provides a feel for what it
looks like on the"ground."
19
33
-
As in 1986, 58 percent of children served at SFSP sites were of
elementary-schoolage. About 25 percent of those served were middle-
or high-school age children;17 percent were preschoolers.
About one-third of sites provided transportation to at least
some children.
A. CHARACTERISTICS OF STATE AGENCIES
In 2001, most states (42) administered the SFSP through their
state education agency, whichalso administers the NSLP and the
School Breakfast Program. In three cases (Michigan,Virginia, and
nonprofit and camp sponsors in New York), the regional offices of
the Food andNutrition Service (FNS), U.S. Department of Agriculture
(USDA), administered the program(Table 11.1).2 In nine states, a
state agency other than the education agency administered
theprogram (including departments of agriculture, health, and
social services).3
SFSP state agencies administered programs that varied widely in
scale. Eight state agencieshad 20 or fewer sponsors, 31 state
agencies had from 21 to 100 sponsors, and 15 had more than100. The
number of sites that the state agencies were responsible for
monitoring also variedwidely. About two-thirds of the states had
between 101 and 1,000 sites; a small group(15 percent) had 100 or
fewer sites, and 12 states (22 percent) had more than 1,000
sites.
The number of SFSP meals that were served in each state varied
widely as well. In 22 states(41 percent), sponsors served 1 million
or fewer meals in the summer of 2001. In most of theother states
(44 percent), sponsors served between 1 million and 4 million
meals. The sponsorsin the state-run portion of the New York program
(the largest "state") served more than12 million meals. On average,
2.4 million SFSP meals were served per state in 2001.
During the late 1970s, as part of efforts to improve program
integrity, state agencies wererequired under federal regulations to
register commercial vendors that provided meals to SFSPsponsors;
the registration process included training vendors on SFSP rules
and inspecting theirfacilities. The federal requirement was dropped
during the late 1990s as part of efforts tosimplify program
administration. Although no longer mandated by FNS, about one-third
of stateagencies (17) still require private vendors to register in
order to be eligible for SFSP contracts.
2In 1986, in contrast, FNS regional offices administered the
SFSP in one-third of the states(Oh ls et al. 1988).
3The two New York agencies (the New York State Department of
Education, whichadministers school and government sponsors, and the
FNS Northeast Regional Office, whichadministers camp and nonprofit
sponsors) are counted separately. Although Wyoming was
notofficially a regional-office-administered state in 2001, the FNS
regional office assisted theWyoming Department of Health. For 2002,
Wyoming has changed its state agency to theDepartment of
Education.
-
TABLE II.1
CHARACTERISTICS OF STATE AGENCIES
Number of StateAgencies
Percentage of StateAgencies
Type of Agenc?State education agency 42 77.8Other state agency 9
16.7FNS regional office 3 5.6
Number of SFSP Sponsors in 2001"1 to 20 8 14.821 to 50 13 24.151
to 100 18 33.3101 to 150 9 16.7151 to 200 3 5.6>201 3 5.6
Median 69Mean 81
Number of SFSP Sites in 2001'1 to 50 4 7.451 to 100 4 7.4101 to
250 13 24.1251 to 500 9 16.7501 to 750 7 13.0751 to 1,000 5
9.31,001 to 1,500 6 11.1>1,500 6 11.1
Median 412Mean 657
21
33
-
TABLE II.1 (continued)
Number of StateAgencies
Percentage of StateAgencies
Number of SFSP Meals Served inSummer 2001'
2 to 4 million 10 18.5>4 to 8 million 4 7.4>8 to 12
million 3 5.6>12 million 1 1.8
Mean (millions) 2.4
Number of States Continuing VendorRegistrations 17 31.5
Total 54
SOURCE: See the footnotes.
aDerived from state plans submitted to FNS and state contact
information.
bTabulated from SFSP Implementation Study, Sponsor-Site Database
(2001).
'Preliminary estimates provided by the Child Nutrition Division,
FNS (January 2002). Thesedata reflect claims reported to FNS by
state agencies for the months of May through September.
"Tabulated from SFSP Implementation Study, State Administrator
Census (2001).
FNS = Food and Nutrition Service.
-
B. CHARACTERISTICS OF SPONSORS
Several sources of data on the characteristics of SFSP sponsors
were collected for this study.The SFSP Sponsor-Site Database,
compiled from lists of sponsors submitted by state
agencies,provides a census of key characteristics. The information
in the database makes it possible toexamine the characteristics of
each major type of sponsor.4 The Sponsor Survey is the source
ofmore-detailed information on sponsors' characteristics.5 Data
from both sources are presentedhere in two different ways: (1) to
reflect the percentage of sponsors nationally with
particularcharacteristics, and (2) to reflect the percentage of
SFSP meals nationally served by sponsorswith particular
characteristics.6
To assess changes in sponsors' characteristics over time, both
FNS program data andcomparisons of the data from the 2001 Sponsor
Survey (or the Sponsor-Site Database) with 1986data from the report
by Oh ls et al. (1988) are used. The FNS data are essentially a
census of theSFSP; however, they were collected only for SFSP
sponsors and sites operating in July, so theyunderstate the size of
the overall program. Comparisons of data from the two surveys must
bemade with caution because of differences in data collection
approaches and sample design in thetwo studies, and because both
sets of estimates are subject to statistical sampling
error;comparisons of the 1986 survey data and the 2001 census data
are more reliable, as only one setof estimates is subject to
sampling error.
I. Overview of Sponsors' Characteristics
About half the 2001 SFSP sponsors were SFAs, and they served
half the SFSP meals(Table 11.2).7 Only 14 percent of sponsors were
government agencies, but these sponsors served31 percent of all
meals in 2001, indicating that government agencies are larger than
the averagesponsor. Nonprofit organizations, National Youth Sports
Programs (NYSPs), and residential
4The database also can be used to examine other subgroups.
Appendix D providestabulations from the database on sponsor
characteristics at the state and regional levels. ChapterIV
compares new and continuing sponsors.
5Appendix D compares the census data from the Sponsor-Site
Database and the SponsorSurvey data on key sponsor characteristics.
As discussed in greater detail in Appendix D, thesurvey data were
weighted to match key control totals from the Sponsor-Site
Database.
6Using the census data, the first type of tabulation is based on
unweighted tabulations of thedata, with each sponsor receiving an
equal weight, regardless of size. For the second type oftabulation,
each sponsor is weighted by the total number of meals it served, so
that largersponsors are weighted more heavily than are smaller
ones. For the survey data, a different set ofsampling weights
corresponds to each type of tabulation, but the underlying idea is
the same.(See Appendix B for additional details on sampling
weights.)
7According to responses to the Sponsor Survey, about 2 percent
of sponsors were privateschools. Thus, private school sponsors
comprise a small portion of all school sponsors.
-
TABLE 11.2
SELECTED CHARACTERISTICS OF SPONSORS
Percentage ofSponsors
StandardError
Percentage ofMeals Served
StandardError
Type of Sponsor'School 48.5 50.9Government 14.2 -
31.4Camp/Upward Bound 16.4 6.8NYSP 3.5 1.4Nonprofit organization
17.5 9.5
New Sponsor' 10.1 2.9
Number of Years as SponsorbFirst year (new this year) 2 (1.0) 2
(1.0)2 to 5 41 (5.6) 16 (4.0)>6 57 (5.6) 82 (4.2)Unknown (but
>1 year) 0 (0.4) 1 (0.7)
Number of Sites Sponsored'1 49.6 10.82 to 5 27.1 12.46 to 10 9.1
8.0llto 50 11.6 24.251 to 100 1.5 9.3101 to 200 0.8 12.4201 to 300
0.2 6.9>300 0.1 16.0
Median 2.0 34.0Mean 8.1 205.5
Average Daily Attendance5,000 2 (0.5) 38 (9.2)
Median 145 (20) 2,026 (1,779)Mean 687 (138) 35,631 (18,459)
24
38
-
TABLE 11.2 (continued)
Percentage ofSponsors
StandardError
Percentage ofMeals Served
StandardError
Total Meals Served DuringSummer'
100,001 4.5 56.7
Median 7,285 153,365Mean 29,858 1,160,433
Duration of Program (CalendarWeeks)
Missing 2.9 1.8
-
TABLE II.2 (continued)
Percentage ofSponsors
StandardError
Percentage ofMeals Served
StandardError
Meals Offered at One or MoreSites'
Breakfast 72.1 78.4Lunch 98.4 99.7Supper 20.7 13.7Any snack 15.2
33.0
Type of Meal Preparation"Self-preparation on site 63 (4.8) 26
(5.5)Self-preparation at central
kitchen 14 (3.1) 16 (4.0)Self-preparation on site or at
central kitchen 5 (1.7) 28 (9.0)SFA as vendor 6 (1.9) 10
(3.4)Private vendor 13 (3.8) 20 (6.1)
Total SponsorsSponsor-SiteDatabase 4,372
Sample SizeSponsor Survey 126
SOURCE: SFSP Implementation Study, Sponsor Survey and
Sponsor-Site Database (2001).
"From the Sponsor-Site Database. These data are a census, so
they have no standard errors. Sponsors withmissing data were
omitted from the tabulations.
"From the Sponsor Survey. Tabulations are weighted to be
representative of sponsors nationally.
`Average daily attendance was reported in the Sponsor Survey.
Sponsors that served more than one meal werenot instructed in any
way on how to calculate average daily attendance, but most seem to
have used eitherattendance at the meal serving the largest number
of children or usual attendance for the activity programassociated
with the meal service. The measure does not adjust for differences
in the number of days thatprograms are open. By contrast, the
measure, total meals, counts every meal served during the
summer.
NYSP = National Youth Sports Program; SFA = School Food
Authority.
26
40
-
camps/Upward Bound programs comprised 37 percent of sponsors but
served only 18 percent ofthe meals.8
Sponsors tend to be stable. As reported in the survey, 57
percent of all sponsors had been inthe SFSP for longer than 6 years
(and these sponsors served 82 percent of all meals). The
listsprovided by the state agencies indicate that 10 percent of the
2001 SFSP sponsors had notsponsored the SFSP in 2000; some of these
sponsors may have offered the program in previousyears. Based on
the Sponsor Survey, only 2 percent of sponsors were new. Some of
the lists ofnew sponsors that were used to select new sponsors for
the survey were incomplete, so it ispossible that new sponsors were
underrepresented somewhat in the survey. In addition, sponsorsthat
had not operated in 2000 but had operated in the past may not have
reported themselves tobe "new."
In 2001, the SFSP had a few very large sponsors, and a large
number of very small ones.This breakdown holds whether measured in
terms of the number of sites, average dailyattendance, or total
meals served during the year.
Most SFSP sponsors operated only a few sites. About half the
sponsors had only 1 site, and86 percent had 10 sites or fewer.
However, sponsors with 10 sites or fewer served fewer thanone-third
of all meals. By contrast, only 1 percent of sponsors operated more
than 100 sites, butthey served 35 percent of all meals.
Another indicator of the wide dispersion in sponsor size is that
most sponsors (85 percent)had daily attendance of 500 children or
fewer, but sponsors that had more than 500 childrenattending per
day served 74 percent of all meals.9 A similar dispersion is
evident in the numberof meals served during the course of the
summer. The majority of sponsors (60 percent) served10,000 meals or
fewer. Although only 5 percent served more than 100,000 meals,
these sponsorsserved 57 percent of all meals served.
Sponsors varied greatly in the duration of their SFSP programs,
but the average programoperated for 7.5 weeks. About one-quarter of
programs were between 4 and 6 weeks in duration,one-quarter were
between 6 and 8 weeks in duration, and one-quarter were between 8
and10 weeks in duration. The small group of sponsors (3 percent)
that operated for 12 weeks orlonger includes sponsors that operated
year-round or almost year-round to serve off-trackchildren in
year-round school districts.
8Upward Bound programs are federally funded educational programs
for disadvantagedyouths, which are operated by colleges or
universities. Although they may be residential ornonresidential,
they are grouped with residential camps in this report, as most
state agencies usethat classification.
9Average daily attendance measures the number of children eating
at any time during theday. This measure counts children who receive
several meals in a day only once; furthermore,the measure does not
adjust for differences in the number of days that programs are
open. Bycontrast, the measure, total meals, counts every meal
served during the summer.
27
41
-
More than half (55 percent) of sponsors operated only open
sites.10 About 11 percent ranonly enrolled sites, and another 13
percent ran a combination of open sites and enrolled sites.
Ingeneral, sponsors that operated both types of sites were large;
they served 24 percent of allmeals. Nineteen percent operated
residential camp sites (including Upward Bound sites), butthey
served only about 8 percent of the meals. These sponsors served a
relatively smallpercentage of meals because they ran fewer sites
than did other sponsors (as discussed in SectionB.2. a).
The majority (56 percent) of sponsors had at least one rural
site, but these sponsors servedjust 30 percent of all meals. Most
sponsors with any rural sites operated only rural sites;53 percent
of all sponsors had only rural sites. Nine percent of sponsors had
one or moremigrant sites. These sites serve children from migrant
worker families and are allowed to servean extra meal. About 8
percent of sponsors reported having one or more sites that
servedhomeless children. Mobile sitessites located on a bus or van
that moves among severallocationshave been developed during the
past few years. Three percent of sponsors reportedoperating at
least one mobile site.
Lunch is typically the primary meal at SFSP sites, and 98
percent of sponsors served thismeal. Nearly three out of four
sponsors (72 percent) had sites that served breakfast, but
breakfastwas not necessarily served at all these sponsors' sites.
About 21 percent of all sponsors had sitesthat served supper, but
these sponsors served only 14 percent of all meals; most sponsors
thatserve supper are residential camps, which generally have only
one site. Fifteen percent ofsponsors ran sites that served snacks;
these sponsors served about one-third of all meals. Somelarger
sponsors served snacks at only a few of their sites.
Nearly two-thirds of sponsors (63 percent) prepared all their
meals at their sites, but thesesponsors served only 26 percent of
all meals." By contrast, 14 percent of sponsors preparedmeals for
all their sites at central kitchens, and 5 percent prepared meals
at some sites anddelivered meals to other sites from a central
kitchen. These two groups together served44 percent of all meals.
Thirteen percent of sponsors obtained meals from private vendors,
and6 percent obtained meals from SFAs (acting as vendors rather
than as sponsors). Sponsors thatused vendors served 30 percent of
meals overall.
mAppendix A describes procedures used for resolving
discrepancies between sponsor andsite data regarding the
open/enrolled status of sites. In general, sponsors' applications
were usedwhen these data sources did not agree.
"See Appendix A for a discussion of cleaning of sponsor and site
data on meal preparation.In general, discrepancies in sponsors' and
site supervisors' reports of meal preparation methodswere resolved
by consulting sponsors' applications. Note also that some meals
prepared incentral kitchens may be warmed up or portioned out at
the sites; for this study, sponsors that usedthis approach have
been classified as using central kitchen preparation.
28
42
-
2. Characteristics of Different Types of Sponsors
The major types of sponsorsSFAs, government agencies,
residential camps, NYSPsponsors, and nonprofit sponsorsoffer very
different types of programs. The Sponsor-SiteDatabase provides a
census of data on sponsors and enables one to consider how each
sponsortype differs for a limited set of characteristics.
Unfortunately, the survey sample is not sufficientto provide
reliable data for each sponsor type. Thus, for some key variables
obtained from thesurvey but not available in the census,
comparisons are between school sponsors and nonschoolsponsors. Much
recent SFSP policy has focused on expanding the role of school
sponsors, sothese comparisons are also of great interest.
a. Characteristics of Major Types of Sponsors, from the Sponsor
Census
In general, different sponsor types served different meals
(Table 11.3). Other than camps,school sponsors were the most likely
group to serve breakfast (75 percent did so), perhapsbecause many
have become accustomed to serving breakfast during the school year.
By contrast,government sponsors were the least likely to serve
breakfast; 47 percent served this meal.Suppers were largely served
at camp sites, but some sponsors in every group served
supper.School sponsors were least likely to serve supper; only 3
percent served this meal.
Camp, NYSP, and nonprofit sponsors never operated large numbers
of sites. In 2001, about85 percent of camp sponsors and about the
same percentage of NYSP sponsors had one site;about three-quarters
of nonprofit sponsors had five or fewer sites (Table 11.3).
Furthermore, noNYSP sponsor had more than 10 sites, and no camp,
Upward Bound, or nonprofit sponsor hadmore than 50.12 Despite their
small number of sites, however, some of these sponsors served
arelatively large number of meals per site: their sites were
relatively large, and most campsponsors served three meals.
Government sponsors generally were the largest sponsors overall,
as they ran an average of21 sites and, on average, served the
largest number of meals. (By contrast, schools ran anaverage of
eight sites, nonprofit organizations, five sites, camps, one site,
and NYSPs, one site.)School sponsors fell in the middle of the size
range, on average; although a substantial fraction(42 percent)
operated only one site, some school sponsors were very large.
Camp and NYSP sponsors operated shorter programs, on average,
than did other types ofsponsors. Camp programs ran for an average
of 6.6 weeks, and NYSPs ran for an average of5.3 weeks. (NYSPs
almost always operated for 4 to 6 weeks.) By contrast, nonprofit
sponsorsoperated programs with the longest average duration (8.9
weeks) and were most likely to operateprograms that lasted 10 weeks
or longer (20 percent did so). The average durations of school-
12As noted in Chapter I, although regulations prevent nonprofit
organizations from operatingmore than 25 sites, FNS may grant
waivers from these regulations. According to the Sponsor-Site
Database, 12 nonprofit sponsors nationally operated between 25 and
50 sites; none operatedmore than 50 sites.
29
-
TABLE 11.3
SELECTED CHARACTERISTICS OF SPONSORS, BY TYPE OF
SPONSOR(Percentages)
School Government Camp NYSP Nonprofit
Meals OfferedBreakfast 75.4 47.0 98.3 66.9 59.6Lunch 98.1 98.6
98.9 95.4 98.0Supper 3.4 7.1 95.4 15.2 10.5Any snack 11.9 22.1 8.1
22.5 24.3
Number of Sites1 42.3 32.1 84.5 85.4 44.22 to 5 32.7 23.4 13.8
13.9 29.66 to 10 10.6 11.1 1.1 0.7 12.611 to 50 12.0 23.7 0.6 0.0
13.651 to 100 1.7 4.7 0.0 0.0 0.0101 to 200 0.6 3.4 0.0 0.0 0.0201
to 300 0.1 1.3 0.0 0.0 0.0>300 0.1 0.5 0.0 0.0 0.0
Median 2.0 4.0 1.0 1.0 2.0Mean 8.2 21.5 1.3 1.2 4.8
Total Meals Served
-
TABLE 11.3 (continued)
School Government Camp NYSP Nonprofit
Duration of Program(Calendar Weeks)
Missing 2.7 0.8 3.8 1.3 4.7
-
sponsored programs and government-sponsored programs fell in the
middle of this range(7.3 weeks and 8.3 weeks, respectively), and
both sponsor types included sponsors with a widerange of
durations.
Ten percent of all sponsors were new in 2001; fully 19 percent
of nonprofit sponsors werenew. About 10 percent of school sponsors
were new (which is proportionate to sponsorsoverall). Smaller
proportions of government, camp, and NYSP sponsors were new.
b. Comparison of School Sponsors and Nonschool Sponsors, from
the Survey Data
Almost all school sponsors (97 percent) prepared their meals
themselves; nonschoolsponsors were much more likely to rely on
vendors (33 percent did so, compared with only3 percent of school
sponsors; see Table 11.4).13 Sixty-eight percent of school sponsors
alwaysprepared meals on site, and 30 percent made some use of
central kitchens. Two-thirds ofnonschool sponsors prepared their
own meals either on site or at a central kitchen. However,22
percent of nonschool sponsors used private vendors, compared with
only 3 percent of schoolsponsors. Eleven percent of nonschool
sponsors used an SFA as a vendor.
School sponsors were substantially more likely than nonschool
sponsors to operate any ruralsites (71 percent versus 42 percent),
to operate only rural sites (68 percent versus 40 percent),and to
operate migrant sites (18 percent versus 2 percent). One possible
explanation is that, inrural areas, there may be few organizations
other than school districts with the ability to serve assponsors.
School and nonschool sponsors did not differ significantly in their
use of mobile sitesor in whether their sites served homeless
children.
Although school sponsors generally were larger than nonschool
sponsors, as measured byaverage daily attendance, the differences
were not statistically significant. About one-quarter ofschool
sponsors and about 36 percent of nonschool sponsors had fewer than
100 children attendper day. About half of both school sponsors and
nonschool sponsors served 100 to 500 childrenper day.
3. Changes in Characteristics of Sponsors Since 1986
The last major study of the SFSP was undertaken in 1986 (Ohls et
al. 1988), and it seemsuseful to assess how SFSP sponsorship has
changed since then. This section compares the twoperiods, using FNS
administrative data, and compares findings from survey and census
dataobtained in the current study with findings from survey data
collected by Ohls et al.
13School sponsors may use private vendors if they do so for the
NSLP and are continuing thesame contract.
32
-
TABLE 11.4
SELECTED CHARACTERISTICS OF SPONSORS,BY SCHOOL/NONSCHOOL
SPONSOR
(Percentage of Sponsors)
SchoolSponsor
StandardError
NonschoolSponsor
StandardError
Type of Meal PreparationSelf-preparation on site 68 (5.7) 58**
(7.6)Self-preparation at central kitchen 18 (4.9) 10
(4.1)Self-preparation on site or at central
kitchen 11 (3.4) 0 (0.0)SFA vendor 0 (0.0) 11 (3.7)Private
vendor 3 (2.7) 22 (6.6)
Presence of Special SitesAny rural sites 71 (6.9) 42** (7.8)All
rural sites 68 (7.2) 40* (7.9)Any migrant sites 18 (5.0) 2**
(1.5)Any mobile sites 3 (1.8) 4 (2.5)Any sites that serve
homeless
children 6 (3.8) 9 (4.0)
Average Daily Attendance5,000 2 (0.9) 1 (0.6)
Median 177 (44) 116 (29)Mean 883 (250) 501 (130)
Sample Size 61 - 64 -SOURCE: SFSP Implementation Study, Sponsor
Survey (2001).
SFA = School Food Authority.
*Significantly different at the .05 level, chi-squared
test.**Significantly different at the .01 level, chi-squared
test.
33
-
a. Program Growth Overall and Among Sponsors of Different
Types
For many years, FNS has collected detailed SFSP program data
f